[1500 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool

[Wandering Albatross]

Much sooner than that.  Some might say it is already bad for small miners.  But I suppose it depends on what you consider "bad for small miners"?  If it takes a typical miner several hours to find a share, that is going to lead to painful variance.

I made a chart to illustrate what happens as the pool gets larger.  The chart is based on an overall bitcoin difficulty of 1.4 million (approx what we are now and will be for the next round).

Thanks for that chart. Definitely important for any small miner. I have a lot to learn here but there seems to be no point at all mining in p2pool if you're a small miner. And small miners will figure this out quickly and move back to their favorite pool.

[forrestv]

my logfile is fast growing 
how I can cut it?

where I can see full run_p2pool.exe options list?

P2Pool will automatically trim the log file if it reaches 100MB, so you shouldn't have to worry about the size of it. To see options, just run "run_p2pool.exe --help".

forrestv; will you consider adding a command line option that limits the total number of connections? My p2pool instance currently has 47 connections, and this plus bitcoind's connections I fear might be too much for my cheap router. At least this could be the case when other people need to use the connection too for torrent downloads etc.

Okay, will do.

The coinbase txn is limited to a certain number of payout (otherwise these p2pool blocks would be ENORMOUS)
So what happens to all the small unpaid amounts?
Since the share-chain only lasts N days ...

The generation transactions pay everyone. The sharechain lasts one day, and there are only a few hundred unique payout addresses which are all paid out.

Thanks for that chart. Definitely important for any small miner. I have a lot to learn here but there seems to be no point at all mining in p2pool if you're a small miner. And small miners will figure this out quickly and move back to their favorite pool.

"No point"? Your expected payout is the same as mining on a pool, but there's just more variance. There are lots of solo miners that are somehow happy with getting less than one block per day, so why can't these small miners cope with less than one share? That said, I'm thinking about several methods that would decrease P2Pool's difficulty, such as letting the big miners on P2Pool voluntarily raise their difficulty so that the smaller miners can use a reduced difficulty while maintaining 10 seconds per share.

[Elmojo]

Hi guys,
I just got P2Pool (kinda) working with Ufaminer on my GTX480.
Now that everything 'appears' to be working correctly, how do i know if it's actually doing useful work that will eventually earn me a payout?
When I was on deepbit with guiminer, at least I could see the shares being created and the small transactions trickling in.
I'm all for the goals of P2Pool, and would like to stick with it, but I'm a little concerned that I'm wasting time and power over here.
Here's a screenshot


Can you guys take a look at it and see if it looks normal?
At the time of this shot, I'd been mining for a little under an hour.
Thanks,
-Elmojo

[ThiagoCMC]

We just hit 260GHash!! 

[ThiagoCMC]

Public P2Pool, how?!

Guys, I would like to setup a Public P2Pool for anyone that doesn't have knowledge / time to setup its own P2Pool node...

How can I do it?!

Thanks!
Thiago

[Elmojo]

Looks normal. The shares (you have one in that screenshot) are what will earn you a payout. The current payout is how much you should expect to earn if the pool finds a block.
P2Pool shares are much more difficult than traditional pool shares, so don't be alarmed. It looks like you should expect a share every 3.9 hours on average.

Great, thanks for checking.
So I shouldn't be concerned that the 1 share I have is a stale?

[kano]

...

The coinbase txn is limited to a certain number of payout (otherwise these p2pool blocks would be ENORMOUS)
So what happens to all the small unpaid amounts?
Since the share-chain only lasts N days ...

The generation transactions pay everyone. The sharechain lasts one day, and there are only a few hundred unique payout addresses which are all paid out.

So then to repeat a question I asked someone before (looks like I better go find who that was)
If the share difficulty is say 500 (which means 2759 shares per block) and they are mined by 1000 people, then the block will have 1000 payments in it?
I hope not - killing the block-chain seems like a bad idea no matter how ideologically good it seems.

Thanks for that chart. Definitely important for any small miner. I have a lot to learn here but there seems to be no point at all mining in p2pool if you're a small miner. And small miners will figure this out quickly and move back to their favorite pool.

"No point"? Your expected payout is the same as mining on a pool, but there's just more variance. There are lots of solo miners that are somehow happy with getting less than one block per day, so why can't these small miners cope with less than one share? That said, I'm thinking about several methods that would decrease P2Pool's difficulty, such as letting the big miners on P2Pool voluntarily raise their difficulty so that the smaller miners can use a reduced difficulty while maintaining 10 seconds per share.

... the reason why pools exist in the first place ...

[Red Emerald]

P2Pool shares are much more difficult than traditional pool shares, so don't be alarmed. It looks like you should expect a share every 3.9 hours on average.

Can we put this in BIG GIANT RED letters somewhere? It seems like every third post is someone asking why its taking so long to get shares...

[Elmojo]

P2Pool shares are much more difficult than traditional pool shares, so don't be alarmed. It looks like you should expect a share every 3.9 hours on average.

Can we put this in BIG GIANT RED letters somewhere? It seems like every third post is someone asking why its taking so long to get shares...

That's a great idea, but i wan't asking why I wasn't getting shares. I was asking if the general activity of my miner (as shown in the screenshot) looked normal, or if it looked like I had a problem.
I do agree that some better documentation and a FAQ would be welcome.

[kano]

I hope not - killing the block-chain seems like a bad idea no matter how ideologically good it seems.

How is it killing the block chain?

Having coinbase txn's something like 40k bytes before adding other txn's ... that explain it?

[jothan]

That's why I have a Python script with a software PID loop that keeps the temperature at 81.0 C +- .5 C on my 2x6990. As soon as I ventilate the room, the fans go down pretty fast. If anybody is interested, I can put that script on GitHub.

Yep, I'd be interested to see that script.

Here it is: https://github.com/jothan/fanatic

Disclaimer: provided as-is; it does not auto-detect GPUs, my 2 6990s are somewhat hardcoded in there. You should be able to find out the adapter numbers for the fans and the temperatures by using the atitweak script in the adl3 python module.

Edit: have a look at the TODO, it is pretty much what is missing to enable more people to use it easely. I will fix some of those items soon.

[ancow]

Thanks for that chart. Definitely important for any small miner. I have a lot to learn here but there seems to be no point at all mining in p2pool if you're a small miner. And small miners will figure this out quickly and move back to their favorite pool.

"No point"? Your expected payout is the same as mining on a pool, but there's just more variance. There are lots of solo miners that are somehow happy with getting less than one block per day, so why can't these small miners cope with less than one share?

The way the difficulty is going, us smaller miners will, at some point, have something like 1 share every couple of days/weeks, and that's the real problem. After all, if we just wanted the block reward + fees, we could try solo-mining and not finding a block for months.

Nvidia miners with ~20 MH/s are already in the ~1 share/day vicinity. With variance that means quite a few days of not getting paid at all, every now and then. With a traditional diff 1 pool, these miners won't get paid much, but at least regularly.
Yes, in the end the balance is the same/similar, but the point of using pools is reduction of variance, not getting paid.

(BTW, not complaining here, just pointing out why p2pool is getting less and less attractive for us smaller miners.)

[ThiagoCMC]

HOLY CRAP!

I FOUND ANOTHER BLOCK!!  ^_^


2012-02-13 23:57:55.617940 GOT BLOCK FROM MINER! Passing to bitcoind! bitcoin: fa33fa88b36259e72b663fbd67f9b756328a0f7c9118aaaacd
2012-02-13 23:57:55.648910 GOT BLOCK FROM PEER! Passing to bitcoind! e0c340a2 bitcoin: fa33fa88b36259e72b663fbd67f9b756328a0f7c9118aaaacd

CHEERS!
Thiago

[ThiagoCMC]

HOLY CRAP!

I FOUND ANOTHER BLOCK!!  ^_^

Save some for the rest of us! Just kidding, find them all! (As long as you are in the same pool as me.) 

hihiHIHIHI ^_^

[jothan]

HOLY CRAP!

I FOUND ANOTHER BLOCK!!  ^_^

Sheesh, are they giving blocks away at tradeshows yet ?

*checks couch for a lost block*

[rjk]

Woot, I just got p2pool running successfully as a Windows service (using srvany and instsrv, and some registry hacking), and loading its config from a file. Life is good.

[Vanderbleek]

Woot, I just got p2pool running successfully as a Windows service (using srvany and instsrv, and some registry hacking), and loading its config from a file. Life is good.

Post a guide?

Or an installer -- might help people who don't like/don't want a command window running.

[rjk]

Woot, I just got p2pool running successfully as a Windows service (using srvany and instsrv, and some registry hacking), and loading its config from a file. Life is good.

Post a guide?

Or an installer -- might help people who don't like/don't want a command window running.

I hacked up the service wrappers that were posted somewhere else on the forum so that I could run bitcoind and namecoind as services and I'm getting ready to post my updates eventually. I'll make a guide at the same time. Might be a few days. I'll see about creating a NSIS based install script, but that will be a bit more difficult.

[Vanderbleek]

Woot, I just got p2pool running successfully as a Windows service (using srvany and instsrv, and some registry hacking), and loading its config from a file. Life is good.

Post a guide?

Or an installer -- might help people who don't like/don't want a command window running.

I hacked up the service wrappers that were posted somewhere else on the forum so that I could run bitcoind and namecoind as services and I'm getting ready to post my updates eventually. I'll make a guide at the same time. Might be a few days. I'll see about creating a NSIS based install script, but that will be a bit more difficult.

Cool, I was just suggesting it if you hadn't thought about it.

[forrestv]

So then to repeat a question I asked someone before (looks like I better go find who that was)
If the share difficulty is say 500 (which means 2759 shares per block) and they are mined by 1000 people, then the block will have 1000 payments in it?
I hope not - killing the block-chain seems like a bad idea no matter how ideologically good it seems.

Yes, though having the hash rate distribution be completely uniform is impossible. Some people will get a lot of the shares, so having one address for every two shares is impossible. In addition, a move to SMPPS with a set minimum payout in the future is possible, though I'm not sure that what we have now will be a problem.